What consitutes a new idea? How do you measure inspiration or a moment of genius? These may sound like silly questions, but policy makers and governments for years have attempted to both monitor and positively impact the rate at which new, original, useful ideas, or innovations take place.
Innovation is thought to be vital for economic growth, prosperity and development. When you think about most of the advancements of the last decades or centuries, many of them can be attributed to a single person or group of people with an unexpected idea. Think of Thomas Edison and the lightbulb, the Wright Brothers and the first airplane.
Many think that innovation will be extremely important in the future, for example to overcome societal challenges and our dependency on unsustainable energy use.
Until now, not terribly well. For decades those studying innovation have used patent applications to measure innovation. Patents are a good way to measure innovation in a number of ways: a patent application takes place in a set moment a time, a patent is a legal claim of a new product, design, or use, which is recorded, so in theory no two patents are the same, or would be recorded twice.
On the other hand…
It costs money to apply for a patent, so people only apply when they need to protect their idea or innovation for commercial purposes. So anytime someone comes up with an innovation that doesn’t need to be protected or commercialized, they would likely not patent it.
With the rise of the internet, alot of innovation is being put out openly, people create programs, and even whole programming languages, and share them freely with the world. So its pretty likely that patents are underestimating innovation. Using patents to measure innovation seems kind of like using articles published in academic journals to measure human knowledge. It is a not a bad place to start, and it is easy to count, but it would exclude everything put out freely online, on Wikipedia for example. So we thought imagining new ways of measuring innovation could be useful.
We wondered to what extent innovation can be measured from looking at people who other people think are innovative online, to see if they could be a better or different way to measure innovation, rather than patents?
We used a website called GitHub, which allows people to show new web development projects openly online, and got all of the counts for how many people were following other people. Thereby we count the relative importance of people´s projects, which we see as an innovation measure.
We thought that these ‘follower numbers’ seemed like a good thing to measure because if people are following another person on GitHub, it means that the person is doing something interesting that people want to see in the future, and expect new or interesting ideas from.
For a range of 135 cities in 16 countries, we take the follower counts for all GitHub users and compare them to traditional patent data. Since we are interested in innovation more generally, we also look at a few other indicators, such as economic output, air pollution, available urban greenspace, and the importance of the city, to find out whether these other indicators affect one or both measures of innovation.
Let´s look at how many GitHub users there are in different countries and compare their follower numbers for the United States, Germany, France and Japan.
INTERACTIVE FIGURE
When looking at the follower counts for these selected countries, we can see that many users have 1-9 or 10-19 followers, but there is a noticeable drop at the threshold of 20. This is how we decided that all the GitHub users who have over 20 followers should be considered innovative.
Let´s now look at how many of those GitHub users with more than 20 followers there are in different cities. The user count depicted in the maps represents users per 10,000 population.
As we can see, there are some GitHub hotspots in Europe, especially in Berlin, Amsterdam, Oslo and Stockholm.
In comparison, cities in Japan show a relatevely low GitHub diffusion with only few users with over 20 followers. One exception is Tokyo.
In America, many cities feature comparatively many GitHub users with over 20 followers, for example in Madison, Portland, Seattle and Austin.
In our research we find a strong relationship between patents and followers, but that relationship largely disapears when we control for other aspects. Hence online collaboration and patent activity do not seem to have a direct link.
In this interactive chart you can compare the patent counts to the GitHub user counts. As you can see, there seems to be a general correlation between the two innovation measures, but it is far from clear. This is why the statistical analysis also found no real direct connection for the whole of our dataset.
INTERACTIVE FIGURE
The interesting part is, however, that both patents and follower numbers are related to air pollution, a city´s economic output, and how important the city is in relation to the country. So, although there might not be a direct link between the two, online collaboration and patent intensity seem to thrive under similar conditions.
It’s hard to tell at this point whether we are measuring the same or a different type of innovation when comparing online collaboration to patents. Still, we do think we may have found a new and interesting way of thinking of innovation and its measurement. While our 20 follower line, or the use of GitHub at all, might not be a perfect way to measure innovation, in the future using collaborative networks and open data may be a much better way than simply relying on exclusive knowledge in the form of patents.